Automated AC winding design

The paper describes an automated method of determining an optimal ac stator winding pattern for use in electrical machines. The method combines a generalised harmonic winding analysis together with a genetic multi-objective optimisation algorithm. The user simply defines the number of stator slots, poles, phases and whether a single or double layer winding is required along with the objective functions that assess the ‘goodness’ of any winding pattern. The paper includes examples of optimised winding layouts for a number of industrial motor designs including balanced and unbalanced fractional-slot windings and in some cases determines a ‘better’ winding pattern than that recommended in the standard winding design manuals. The flexibility of the approach is further demonstrated with examples of winding patterns for multi-phase and concentrated-tooth windings.